{
  "best_global_step": 17000,
  "best_metric": 0.9708837837837838,
  "best_model_checkpoint": "/kaggle/working/arbert_submeter_training/checkpoints/checkpoint-17000",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 17346,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08648274669203494,
      "grad_norm": 25.747224807739258,
      "learning_rate": 4.856162804104693e-05,
      "loss": 1.2986,
      "step": 500
    },
    {
      "epoch": 0.08648274669203494,
      "eval_accuracy": 0.765245945945946,
      "eval_loss": 0.756655216217041,
      "eval_macro_f1": 0.2987486877044399,
      "eval_runtime": 418.3804,
      "eval_samples_per_second": 884.363,
      "eval_steps_per_second": 3.456,
      "eval_weighted_f1": 0.7565798837380349,
      "step": 500
    },
    {
      "epoch": 0.17296549338406988,
      "grad_norm": 7.156609535217285,
      "learning_rate": 4.7120373573158074e-05,
      "loss": 0.5963,
      "step": 1000
    },
    {
      "epoch": 0.17296549338406988,
      "eval_accuracy": 0.8766405405405405,
      "eval_loss": 0.40730020403862,
      "eval_macro_f1": 0.4197539556932673,
      "eval_runtime": 418.1054,
      "eval_samples_per_second": 884.944,
      "eval_steps_per_second": 3.458,
      "eval_weighted_f1": 0.8676584583901654,
      "step": 1000
    },
    {
      "epoch": 0.25944824007610484,
      "grad_norm": 6.3231658935546875,
      "learning_rate": 4.567911910526923e-05,
      "loss": 0.3789,
      "step": 1500
    },
    {
      "epoch": 0.25944824007610484,
      "eval_accuracy": 0.9137351351351352,
      "eval_loss": 0.2946445643901825,
      "eval_macro_f1": 0.47816321808462303,
      "eval_runtime": 418.8729,
      "eval_samples_per_second": 883.323,
      "eval_steps_per_second": 3.452,
      "eval_weighted_f1": 0.9066607985127312,
      "step": 1500
    },
    {
      "epoch": 0.34593098676813977,
      "grad_norm": 9.21304988861084,
      "learning_rate": 4.423786463738038e-05,
      "loss": 0.2838,
      "step": 2000
    },
    {
      "epoch": 0.34593098676813977,
      "eval_accuracy": 0.9314702702702703,
      "eval_loss": 0.23459294438362122,
      "eval_macro_f1": 0.5072222949676094,
      "eval_runtime": 418.1997,
      "eval_samples_per_second": 884.745,
      "eval_steps_per_second": 3.458,
      "eval_weighted_f1": 0.9263247220377161,
      "step": 2000
    },
    {
      "epoch": 0.4324137334601747,
      "grad_norm": 6.399903297424316,
      "learning_rate": 4.279661016949153e-05,
      "loss": 0.2401,
      "step": 2500
    },
    {
      "epoch": 0.4324137334601747,
      "eval_accuracy": 0.9421918918918919,
      "eval_loss": 0.20360097289085388,
      "eval_macro_f1": 0.5251223211604551,
      "eval_runtime": 418.9782,
      "eval_samples_per_second": 883.101,
      "eval_steps_per_second": 3.451,
      "eval_weighted_f1": 0.9356836777939385,
      "step": 2500
    },
    {
      "epoch": 0.5188964801522097,
      "grad_norm": 7.528863906860352,
      "learning_rate": 4.135535570160268e-05,
      "loss": 0.2168,
      "step": 3000
    },
    {
      "epoch": 0.5188964801522097,
      "eval_accuracy": 0.947972972972973,
      "eval_loss": 0.1820368617773056,
      "eval_macro_f1": 0.5299222018041627,
      "eval_runtime": 418.5666,
      "eval_samples_per_second": 883.969,
      "eval_steps_per_second": 3.455,
      "eval_weighted_f1": 0.9417715175152749,
      "step": 3000
    },
    {
      "epoch": 0.6053792268442446,
      "grad_norm": 3.640698194503784,
      "learning_rate": 3.9914101233713826e-05,
      "loss": 0.1955,
      "step": 3500
    },
    {
      "epoch": 0.6053792268442446,
      "eval_accuracy": 0.9518621621621621,
      "eval_loss": 0.17032282054424286,
      "eval_macro_f1": 0.5375603114820035,
      "eval_runtime": 418.6948,
      "eval_samples_per_second": 883.699,
      "eval_steps_per_second": 3.454,
      "eval_weighted_f1": 0.9454971013709563,
      "step": 3500
    },
    {
      "epoch": 0.6918619735362795,
      "grad_norm": 5.204596996307373,
      "learning_rate": 3.847284676582498e-05,
      "loss": 0.1782,
      "step": 4000
    },
    {
      "epoch": 0.6918619735362795,
      "eval_accuracy": 0.9557081081081081,
      "eval_loss": 0.15780550241470337,
      "eval_macro_f1": 0.5498663380782979,
      "eval_runtime": 418.4036,
      "eval_samples_per_second": 884.314,
      "eval_steps_per_second": 3.456,
      "eval_weighted_f1": 0.9495774047836557,
      "step": 4000
    },
    {
      "epoch": 0.7783447202283145,
      "grad_norm": 2.9452965259552,
      "learning_rate": 3.7031592297936126e-05,
      "loss": 0.1699,
      "step": 4500
    },
    {
      "epoch": 0.7783447202283145,
      "eval_accuracy": 0.9571459459459459,
      "eval_loss": 0.15344275534152985,
      "eval_macro_f1": 0.5465995200000592,
      "eval_runtime": 418.0876,
      "eval_samples_per_second": 884.982,
      "eval_steps_per_second": 3.459,
      "eval_weighted_f1": 0.9507761247869798,
      "step": 4500
    },
    {
      "epoch": 0.8648274669203494,
      "grad_norm": 5.704383850097656,
      "learning_rate": 3.559033783004728e-05,
      "loss": 0.1622,
      "step": 5000
    },
    {
      "epoch": 0.8648274669203494,
      "eval_accuracy": 0.9585405405405405,
      "eval_loss": 0.1496480107307434,
      "eval_macro_f1": 0.5578920159472586,
      "eval_runtime": 418.0713,
      "eval_samples_per_second": 885.017,
      "eval_steps_per_second": 3.459,
      "eval_weighted_f1": 0.9523043519198916,
      "step": 5000
    },
    {
      "epoch": 0.9513102136123843,
      "grad_norm": 3.2097666263580322,
      "learning_rate": 3.4149083362158425e-05,
      "loss": 0.1569,
      "step": 5500
    },
    {
      "epoch": 0.9513102136123843,
      "eval_accuracy": 0.9602459459459459,
      "eval_loss": 0.14497891068458557,
      "eval_macro_f1": 0.5471822620722766,
      "eval_runtime": 418.9758,
      "eval_samples_per_second": 883.106,
      "eval_steps_per_second": 3.451,
      "eval_weighted_f1": 0.9536733224654437,
      "step": 5500
    },
    {
      "epoch": 1.0377064775577272,
      "grad_norm": 4.706158638000488,
      "learning_rate": 3.270782889426958e-05,
      "loss": 0.1422,
      "step": 6000
    },
    {
      "epoch": 1.0377064775577272,
      "eval_accuracy": 0.9593081081081081,
      "eval_loss": 0.1468845158815384,
      "eval_macro_f1": 0.5689795604124087,
      "eval_runtime": 420.1108,
      "eval_samples_per_second": 880.72,
      "eval_steps_per_second": 3.442,
      "eval_weighted_f1": 0.9533441467822756,
      "step": 6000
    },
    {
      "epoch": 1.1241892242497622,
      "grad_norm": 3.167544364929199,
      "learning_rate": 3.1266574426380724e-05,
      "loss": 0.1251,
      "step": 6500
    },
    {
      "epoch": 1.1241892242497622,
      "eval_accuracy": 0.9628189189189189,
      "eval_loss": 0.13693760335445404,
      "eval_macro_f1": 0.5800974551510764,
      "eval_runtime": 420.0357,
      "eval_samples_per_second": 880.877,
      "eval_steps_per_second": 3.443,
      "eval_weighted_f1": 0.9568010100001315,
      "step": 6500
    },
    {
      "epoch": 1.210671970941797,
      "grad_norm": 2.2167930603027344,
      "learning_rate": 2.982531995849187e-05,
      "loss": 0.123,
      "step": 7000
    },
    {
      "epoch": 1.210671970941797,
      "eval_accuracy": 0.9632054054054054,
      "eval_loss": 0.1349489837884903,
      "eval_macro_f1": 0.5686206184279916,
      "eval_runtime": 419.1921,
      "eval_samples_per_second": 882.65,
      "eval_steps_per_second": 3.449,
      "eval_weighted_f1": 0.9570982563798056,
      "step": 7000
    },
    {
      "epoch": 1.297154717633832,
      "grad_norm": 3.8189609050750732,
      "learning_rate": 2.8384065490603024e-05,
      "loss": 0.1208,
      "step": 7500
    },
    {
      "epoch": 1.297154717633832,
      "eval_accuracy": 0.963927027027027,
      "eval_loss": 0.13370193541049957,
      "eval_macro_f1": 0.5746596478621987,
      "eval_runtime": 418.7013,
      "eval_samples_per_second": 883.685,
      "eval_steps_per_second": 3.454,
      "eval_weighted_f1": 0.9576970198616988,
      "step": 7500
    },
    {
      "epoch": 1.3836374643258669,
      "grad_norm": 4.8342461585998535,
      "learning_rate": 2.694281102271417e-05,
      "loss": 0.1188,
      "step": 8000
    },
    {
      "epoch": 1.3836374643258669,
      "eval_accuracy": 0.964472972972973,
      "eval_loss": 0.13066376745700836,
      "eval_macro_f1": 0.5725448949887878,
      "eval_runtime": 418.7971,
      "eval_samples_per_second": 883.483,
      "eval_steps_per_second": 3.453,
      "eval_weighted_f1": 0.9581300625445718,
      "step": 8000
    },
    {
      "epoch": 1.470120211017902,
      "grad_norm": 3.3662569522857666,
      "learning_rate": 2.5501556554825323e-05,
      "loss": 0.1201,
      "step": 8500
    },
    {
      "epoch": 1.470120211017902,
      "eval_accuracy": 0.9651378378378378,
      "eval_loss": 0.12952136993408203,
      "eval_macro_f1": 0.5705480082509794,
      "eval_runtime": 418.5068,
      "eval_samples_per_second": 884.095,
      "eval_steps_per_second": 3.455,
      "eval_weighted_f1": 0.9587555599089429,
      "step": 8500
    },
    {
      "epoch": 1.556602957709937,
      "grad_norm": 3.425612211227417,
      "learning_rate": 2.406030208693647e-05,
      "loss": 0.118,
      "step": 9000
    },
    {
      "epoch": 1.556602957709937,
      "eval_accuracy": 0.965654054054054,
      "eval_loss": 0.1270706057548523,
      "eval_macro_f1": 0.5691587584053437,
      "eval_runtime": 419.0634,
      "eval_samples_per_second": 882.921,
      "eval_steps_per_second": 3.451,
      "eval_weighted_f1": 0.9593746302255395,
      "step": 9000
    },
    {
      "epoch": 1.6430857044019718,
      "grad_norm": 1.4031856060028076,
      "learning_rate": 2.261904761904762e-05,
      "loss": 0.1167,
      "step": 9500
    },
    {
      "epoch": 1.6430857044019718,
      "eval_accuracy": 0.9661621621621621,
      "eval_loss": 0.125877246260643,
      "eval_macro_f1": 0.5795209016258601,
      "eval_runtime": 418.9995,
      "eval_samples_per_second": 883.056,
      "eval_steps_per_second": 3.451,
      "eval_weighted_f1": 0.9600300931445321,
      "step": 9500
    },
    {
      "epoch": 1.7295684510940066,
      "grad_norm": 2.918519973754883,
      "learning_rate": 2.117779315115877e-05,
      "loss": 0.1133,
      "step": 10000
    },
    {
      "epoch": 1.7295684510940066,
      "eval_accuracy": 0.9670945945945946,
      "eval_loss": 0.12201372534036636,
      "eval_macro_f1": 0.5803247802424795,
      "eval_runtime": 418.9179,
      "eval_samples_per_second": 883.228,
      "eval_steps_per_second": 3.452,
      "eval_weighted_f1": 0.9608891715468071,
      "step": 10000
    },
    {
      "epoch": 1.8160511977860416,
      "grad_norm": 2.741220712661743,
      "learning_rate": 1.973653868326992e-05,
      "loss": 0.1116,
      "step": 10500
    },
    {
      "epoch": 1.8160511977860416,
      "eval_accuracy": 0.9673864864864865,
      "eval_loss": 0.12097315490245819,
      "eval_macro_f1": 0.5834823701825592,
      "eval_runtime": 418.7157,
      "eval_samples_per_second": 883.654,
      "eval_steps_per_second": 3.453,
      "eval_weighted_f1": 0.9612394940368674,
      "step": 10500
    },
    {
      "epoch": 1.9025339444780767,
      "grad_norm": 2.6713507175445557,
      "learning_rate": 1.8295284215381068e-05,
      "loss": 0.1103,
      "step": 11000
    },
    {
      "epoch": 1.9025339444780767,
      "eval_accuracy": 0.9684972972972973,
      "eval_loss": 0.11646272242069244,
      "eval_macro_f1": 0.5835802563966344,
      "eval_runtime": 418.3811,
      "eval_samples_per_second": 884.361,
      "eval_steps_per_second": 3.456,
      "eval_weighted_f1": 0.9622871010076606,
      "step": 11000
    },
    {
      "epoch": 1.9890166911701117,
      "grad_norm": 2.58137845993042,
      "learning_rate": 1.6854029747492218e-05,
      "loss": 0.1093,
      "step": 11500
    },
    {
      "epoch": 1.9890166911701117,
      "eval_accuracy": 0.9683432432432433,
      "eval_loss": 0.11723459511995316,
      "eval_macro_f1": 0.5835321485745915,
      "eval_runtime": 419.3158,
      "eval_samples_per_second": 882.39,
      "eval_steps_per_second": 3.448,
      "eval_weighted_f1": 0.9623140045622489,
      "step": 11500
    },
    {
      "epoch": 2.0754129551154543,
      "grad_norm": 3.1163055896759033,
      "learning_rate": 1.5412775279603368e-05,
      "loss": 0.0899,
      "step": 12000
    },
    {
      "epoch": 2.0754129551154543,
      "eval_accuracy": 0.9689324324324324,
      "eval_loss": 0.11741844564676285,
      "eval_macro_f1": 0.5972224837605217,
      "eval_runtime": 419.245,
      "eval_samples_per_second": 882.539,
      "eval_steps_per_second": 3.449,
      "eval_weighted_f1": 0.9632448268921581,
      "step": 12000
    },
    {
      "epoch": 2.1618957018074894,
      "grad_norm": 2.067152738571167,
      "learning_rate": 1.3971520811714517e-05,
      "loss": 0.0851,
      "step": 12500
    },
    {
      "epoch": 2.1618957018074894,
      "eval_accuracy": 0.9692378378378378,
      "eval_loss": 0.11705382913351059,
      "eval_macro_f1": 0.5857631212432319,
      "eval_runtime": 419.6262,
      "eval_samples_per_second": 881.737,
      "eval_steps_per_second": 3.446,
      "eval_weighted_f1": 0.9632093509253533,
      "step": 12500
    },
    {
      "epoch": 2.2483784484995244,
      "grad_norm": 1.862592339515686,
      "learning_rate": 1.2530266343825667e-05,
      "loss": 0.0862,
      "step": 13000
    },
    {
      "epoch": 2.2483784484995244,
      "eval_accuracy": 0.9688189189189189,
      "eval_loss": 0.11716984957456589,
      "eval_macro_f1": 0.590216690584949,
      "eval_runtime": 418.913,
      "eval_samples_per_second": 883.238,
      "eval_steps_per_second": 3.452,
      "eval_weighted_f1": 0.9634743085321129,
      "step": 13000
    },
    {
      "epoch": 2.3348611951915594,
      "grad_norm": 2.883910655975342,
      "learning_rate": 1.1089011875936815e-05,
      "loss": 0.0835,
      "step": 13500
    },
    {
      "epoch": 2.3348611951915594,
      "eval_accuracy": 0.9696405405405405,
      "eval_loss": 0.11618903279304504,
      "eval_macro_f1": 0.5970308633746413,
      "eval_runtime": 418.887,
      "eval_samples_per_second": 883.293,
      "eval_steps_per_second": 3.452,
      "eval_weighted_f1": 0.9638159327084403,
      "step": 13500
    },
    {
      "epoch": 2.421343941883594,
      "grad_norm": 0.7860650420188904,
      "learning_rate": 9.647757408047965e-06,
      "loss": 0.0854,
      "step": 14000
    },
    {
      "epoch": 2.421343941883594,
      "eval_accuracy": 0.9694459459459459,
      "eval_loss": 0.1158711388707161,
      "eval_macro_f1": 0.592536606494846,
      "eval_runtime": 419.3296,
      "eval_samples_per_second": 882.361,
      "eval_steps_per_second": 3.448,
      "eval_weighted_f1": 0.9638996969783739,
      "step": 14000
    },
    {
      "epoch": 2.507826688575629,
      "grad_norm": 3.2249786853790283,
      "learning_rate": 8.206502940159114e-06,
      "loss": 0.0797,
      "step": 14500
    },
    {
      "epoch": 2.507826688575629,
      "eval_accuracy": 0.9697621621621622,
      "eval_loss": 0.11566793918609619,
      "eval_macro_f1": 0.5934226559634111,
      "eval_runtime": 420.7367,
      "eval_samples_per_second": 879.41,
      "eval_steps_per_second": 3.437,
      "eval_weighted_f1": 0.964013311849684,
      "step": 14500
    },
    {
      "epoch": 2.594309435267664,
      "grad_norm": 4.085644245147705,
      "learning_rate": 6.765248472270265e-06,
      "loss": 0.0819,
      "step": 15000
    },
    {
      "epoch": 2.594309435267664,
      "eval_accuracy": 0.9698027027027027,
      "eval_loss": 0.1152041032910347,
      "eval_macro_f1": 0.6003991642644262,
      "eval_runtime": 419.1122,
      "eval_samples_per_second": 882.818,
      "eval_steps_per_second": 3.45,
      "eval_weighted_f1": 0.9643335590541979,
      "step": 15000
    },
    {
      "epoch": 2.680792181959699,
      "grad_norm": 2.286464214324951,
      "learning_rate": 5.323994004381414e-06,
      "loss": 0.0813,
      "step": 15500
    },
    {
      "epoch": 2.680792181959699,
      "eval_accuracy": 0.9702864864864865,
      "eval_loss": 0.1135055348277092,
      "eval_macro_f1": 0.5963674935323597,
      "eval_runtime": 419.7981,
      "eval_samples_per_second": 881.376,
      "eval_steps_per_second": 3.445,
      "eval_weighted_f1": 0.9645615895626319,
      "step": 15500
    },
    {
      "epoch": 2.7672749286517337,
      "grad_norm": 2.7449326515197754,
      "learning_rate": 3.882739536492564e-06,
      "loss": 0.0792,
      "step": 16000
    },
    {
      "epoch": 2.7672749286517337,
      "eval_accuracy": 0.9706054054054054,
      "eval_loss": 0.1119714081287384,
      "eval_macro_f1": 0.6000477857320852,
      "eval_runtime": 419.5757,
      "eval_samples_per_second": 881.843,
      "eval_steps_per_second": 3.446,
      "eval_weighted_f1": 0.9649497278700637,
      "step": 16000
    },
    {
      "epoch": 2.853757675343769,
      "grad_norm": 1.0509530305862427,
      "learning_rate": 2.441485068603713e-06,
      "loss": 0.0786,
      "step": 16500
    },
    {
      "epoch": 2.853757675343769,
      "eval_accuracy": 0.9707135135135135,
      "eval_loss": 0.11158089339733124,
      "eval_macro_f1": 0.6009545374695694,
      "eval_runtime": 418.9418,
      "eval_samples_per_second": 883.178,
      "eval_steps_per_second": 3.452,
      "eval_weighted_f1": 0.9653056586898462,
      "step": 16500
    },
    {
      "epoch": 2.940240422035804,
      "grad_norm": 0.9345380663871765,
      "learning_rate": 1.0002306007148624e-06,
      "loss": 0.0789,
      "step": 17000
    },
    {
      "epoch": 2.940240422035804,
      "eval_accuracy": 0.9708837837837838,
      "eval_loss": 0.11126039922237396,
      "eval_macro_f1": 0.6020869942813024,
      "eval_runtime": 419.4331,
      "eval_samples_per_second": 882.143,
      "eval_steps_per_second": 3.448,
      "eval_weighted_f1": 0.9654166847298334,
      "step": 17000
    }
  ],
  "logging_steps": 500,
  "max_steps": 17346,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.302829698913882e+16,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": null
}